Manufacturing intelligence for early warning of key equipment excursion for advanced equipment control in semiconductor manufacturing

Abstract

As feature sizes of integrated circuits are continuously shrinking in nanotechnologies, mining potentially useful information to extract manufacturing intelligence from big data automatically collected in the wafer fabrication facilities to assist in real time decisions for yield enhancement has become practically crucial to maintain competitive advantages and support intelligent manufacturing for operational excellence. Motivated by real needs, this study aims to develop an effective approach to extract manufacturing intelligence for early detection of key equipment excursion for advanced equipment control to enhance yield and reduce potential loss. For validation, an empirical study was conducted in a leading semiconductor manufacturing company to validate the proposed approach in the developed "early warning system" of newly released equipment to reduce tool excursion and abnormal yield loss. The results have demonstrated practical viability of the proposed approach. Indeed, the developed solution has been implemented in this company.